NRI:INT: Ad-Hoc Collaborative Human-Robot Swarms

点击翻译按钮获取中文摘要

This project addresses the challenge of creating ad-hoc collaborative teams (swarms) of robots and people, where the people are not robotics experts and are interacting with the robots to accomplish an ad-hoc task such as emergency evacuation. The project investigates what the robots should look like, how they should behave, and how they should interact with people without prior experience with robots, to make sure the overall task is accomplished. The scientific insights of this project can apply to different tasks such as crowd control and large scale search missions. Here the research is grounded in a series of room and building evacuation scenarios where robots work autonomously, provide information and guidance, and respond to people's requests and behaviors. This project will impact three graduate classes taught by the researchers, will support undergraduate and masters projects and the results will be incorporated into activities targeted at increasing the participation of groups traditionally underrepresented in engineering. This project advances knowledge in three subfields of robotics and especially at their intersection; swarm design, autonomy, and human-swarm interaction. The challenge includes development of hardware platforms that can traverse and modify their environment and permit safe human-robot interaction, while remaining inexpensive, robust, and low maintenance. Specifically, this will involve blimps to facilitate and provide broad situational awareness by projecting images on the ground, and ground robots with inflatable, interactive bladders. For autonomy, this project considers a top-down approach where individual robot control is synthesized from a high-level task. In addition, algorithms will be developed that enable robots to automatically modify their behavior and provide intuitive feedback based on interactions with humans, the humans' behaviors, and the observed environments. This project will contribute to the field of Human-Robot Interaction by designing and evaluating distributed interactions where the role of the human fluidly changes between being a leader providing instructions to the robots, a cooperative swarm member guided by the robots, an uncooperative member who acts in opposition to the swarm guidance, and a passive member who needs to be physically assisted. Evaluating the human-swarm behavior using established metrics in a variety of increasingly complex environments will provide unique insights into the opportunities and challenges of creating ad-hoc human-robot swarms.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.